Polysomnography (PSG) is considered the gold standard in obstructive sleep apnea-hypopnea syndrome (OSAS) diagnostics, but its availability is still limited. Thus, it seems useful to assess patients pre-diagnostic risk for OSAS to prioritize the use of this examination. The purpose of this study was to assess positive (PPV) and negative (NPV) predictive values of the STOP BANG questionnaire (SBQ) in patients with presumptive diagnosis of OSAS. From a database of 1,171 (880 men) patients of a university based sleep center, 1,123 (847 men) met eligibility criteria and their SBQ scores were subject to the Bayesian analysis. The analysis of PPV and NPV was conducted at all values of SBQ for all subjects, but also separately for males and females, and for total sleep time (TS) and for sleep in the lateral position (LP). The probability of OSAS (AHI ≥ 5) and at least moderate OSAS (AHI ≥ 15) for TS was 0.766 and 0.516, while for LP the values were 0.432 and 0.289, respectively. Overall, due to low specificity, SBQ had low PPV for TS and LP. Negative test result (SBQ < 3) revealed NPV of 0.620 at AHI < 5 and 0.859 at AHI < 15 for TS, while in LP NPV values were 0.935 at AHI < 5 and 1.0 at AHI < 15, (n = 31), while SBQ < 4 generated NPV of 0.943 in LP (n = 105). SBQ did not change probabilities of OSAS to confirm or rebut diagnosis for TS. However, it is highly probable that SQB can rule out OSAS diagnosis at AHI ≥ 15 for LP.